Skip to content

feat: skip re-downloading models on shared storage #541

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
Kangyan-Zhou wants to merge 5 commits into
base: main
Choose a base branch
from
Draft

feat: skip re-downloading models on shared storage #541

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
8e81b16
feat: skip re-downloading models on shared storage
Kangyan-Zhou Mar 16, 2026
3b1140e
feat: extend shared-storage skip to diffusion pipelines and single-fi…
Kangyan-Zhou Apr 3, 2026
8d0baf6
feat: add K8s Lease-based leader election for shared storage downloads
Kangyan-Zhou Apr 3, 2026
256c2e7
fix: handle non-component entries in diffusion model_index.json
Kangyan-Zhou Apr 3, 2026
2f48286
feat: add download timeout to prevent stuck workers
Kangyan-Zhou Apr 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions cmd/model-agent/main.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,6 +43,7 @@ type config struct {
numDownloadWorker int
namespace string
logLevel string
downloadTimeout time.Duration
}

// Logger type alias for zap.SugaredLogger
Expand Down Expand Up @@ -73,6 +74,7 @@ func init() {
rootCmd.PersistentFlags().IntVar(&cfg.numDownloadWorker, "num-download-worker", 5, "Number of download workers")
rootCmd.PersistentFlags().StringVar(&cfg.namespace, "namespace", "ome", "Kubernetes namespace to use")
rootCmd.PersistentFlags().StringVar(&cfg.logLevel, "log-level", "info", "Log level (debug, info, warn, error)")
rootCmd.PersistentFlags().DurationVar(&cfg.downloadTimeout, "download-timeout", 6*time.Hour, "Maximum time allowed for a single model download before it is cancelled")

_ = v.BindPFlags(rootCmd.PersistentFlags())
v.SetEnvKeyReplacer(strings.NewReplacer("-", "_"))
Expand Down Expand Up @@ -268,6 +270,9 @@ func initializeComponents(
logger,
baseModelInformer.Lister(),
clusterBaseModelInformer.Lister(),
cfg.nodeName,
cfg.namespace,
cfg.downloadTimeout,
)
if err != nil {
return nil, nil, fmt.Errorf("failed to create gopher: %w", err)
Expand Down
3 changes: 3 additions & 0 deletions cmd/model-agent/main_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ import (
"net/http/httptest"
"os"
"testing"
"time"

"github.com/prometheus/client_golang/prometheus"
"github.com/spf13/cobra"
Expand Down Expand Up @@ -113,6 +114,7 @@ func TestDefaultConfig(t *testing.T) {
testCmd.Flags().StringVar(&cfg.downloadAuthType, "download-auth-type", "instance-principal", "authentication method for model download")
testCmd.Flags().IntVar(&cfg.numDownloadWorker, "num-download-worker", 3, "number of download workers")
testCmd.Flags().StringVar(&cfg.namespace, "namespace", "ome", "the namespace of the ome model agents daemon set")
testCmd.Flags().DurationVar(&cfg.downloadTimeout, "download-timeout", 6*time.Hour, "maximum time for a single model download")

// Call initConfig to set cfg.nodeName
initConfig(nil, nil)
Expand All @@ -127,6 +129,7 @@ func TestDefaultConfig(t *testing.T) {
assert.Equal(t, "instance-principal", cfg.downloadAuthType)
assert.Equal(t, 3, cfg.numDownloadWorker)
assert.Equal(t, "ome", cfg.namespace)
assert.Equal(t, 6*time.Hour, cfg.downloadTimeout)
}

func TestInitializeLogger(t *testing.T) {
Expand Down
Loading
Loading